Decentralized autonomous organizations promise governance without hierarchy. Token holders vote. Smart contracts execute. No central authority dictates outcomes. That is the marketing.

The data tells a different story. In Market Power and Governance Power, I examine empirical research on governance token distribution in leading DAOs and find Gini coefficients of 0.90 to 0.98 — near-total concentration of voting power at the top. Organizations marketed as decentralized are, by this measure, more concentrated than most traditional corporations.

The delegation mechanism makes it worse. Many DAOs allow token holders to delegate their votes to visible “delegates” — essentially proto-board-members who accumulate voting power from passive holders. The mechanism is structurally identical to the proxy system in corporate governance, where management routinely wins because retail shareholders delegate rather than vote. Formal decentralization produces functional oligarchy.

This matters for antitrust. Regulators cannot take “decentralized” labels at face value. I propose a dual-metric framework that integrates traditional market concentration measures (HHI) with governance concentration metrics (the Nakamoto coefficient, which counts the minimum number of entities needed to reach 51% of governance power). A DAO with a Nakamoto coefficient of 3 is not meaningfully decentralized, regardless of how many token holders appear on the ledger.

The four-quadrant matrix I develop — crossing market concentration with governance concentration — determines when intervention is warranted. Some entities that look competitive by market share are dangerously concentrated by governance. Some that look monopolistic by market share are genuinely decentralized by governance. Antitrust analysis that ignores the governance dimension will get the answer wrong in both directions.

The Beanstalk attack illustrates the stakes. In April 2022, a flash loan attacker borrowed enough governance tokens to pass a malicious proposal, drain $182 million from the protocol treasury, repay the loan, and exit — all in a single thirteen-second Ethereum block. A protocol with a Nakamoto coefficient of 1 (for those thirteen seconds) was marketed as decentralized. The label was the attack surface.

Read the full article: Market Power and Governance Power: New Tools for Antitrust Enforcement in the Decentralized Gig Economy, Competition Policy International (2025).